Complex frequency-domain analysis of sensor response function for quantitative odor identification

In the process of classifier design and training for quantitative odor identification, it is often unavoidable that the odor concentration varies instantaneously with the odor extraction and air replenishment when the volume of sealed sample vessel is fixed. If these changes were ignored, the accuracy of odor classification and quantitative identification will be seriously affected. Therefore, the influences that odor extraction and air replenishment of electronic nose results in concentration change of test sample in the sealed vessel of fixed volume is analyzed in this paper, and the model of concentration varying with time was established firstly. Then, the response function models of olfactory sensor in different concentrations of the same odor were built based on complex frequency-domain analysis method, and the parameters of each model were identified. Finally, based on the features selected from the model parameters, a non-linear correlation model between sensor response and odor concentration was established to provide the basis for quantitative odor identification and mixed odors separation.